What happened
Ant International, the overseas arm of Ant Group, announced a major product upgrade AI agents across its entire suite of global financial services. The initiative covers cross-border payments (Alipay+), merchant services (Antom), treasury (WorldFirst), and digital credit (Bettr). The company introduced proprietary AI systems, including the Antom 3-in-1 for transaction assessment and FalconTST for cash-flow forecasting, alongside new security protocols like 'know-your-agent' (KYA) and AgentSafePay to manage risks associated with autonomous agent transactions.
Ant International has launched what it describes as its largest-ever product upgrade, AI-native solutions across all its major platforms: Alipay+, Antom, WorldFirst, and Bettr. The global rollout has begun and is scheduled to extend through the winter season. The upgrade spans nearly 100 products, including AI assistants for payments, account management, cash-flow forecasting, currency hedging, and global payouts.
The company relies on two proprietary systems for this overhaul. The first is the Antom 3-in-1 , a payment that processes structured records, behavioral sequences, and relationships between users, merchants, and devices to assess transactions. The second is FalconTST, a forecasting model for cash flows and currency exposures. Ant International claims FalconTST has achieved over 93% forecasting accuracy and reduced corporate currency hedging and capital allocation costs by 30 to 60% in real-world applications.
To address security risks associated with autonomous transactions, Ant International introduced a 'know-your-agent' (KYA) framework linked to its Agentic Mobile Protocol. This protocol connects mobile wallet networks to agentic transactions. The security layer includes AgentSafePay, which offers a fund-back guarantee against losses from agent-specific risks, and the Antom Shield framework to guard against AI-driven fraud. The company is also partnering with Visa and Mastercard to establish global agent verification standards.
Gary Liu, CEO of Antom, stated that 89.5% of Antom’s merchants already use AI agents, with 81.4% of payment tasks receiving AI assistance. He emphasized that agent authorization should depend on the task and spending limits rather than job titles. Peng Yang, CEO of Ant International, said the shift prepares global commerce for a 'fully agentised future' and requires a 'future-ready trust infrastructure' paired with financial AI models.
Why it matters
This move represents a significant shift in how global financial infrastructure is being adapted for autonomous AI agents. By integrating AI directly into core banking and payment functions, Ant International is attempting to standardize how agents interact with financial systems, including verification and liability frameworks. This is critical for the broader adoption of agentic commerce, as it addresses the trust and security gaps that currently hinder widespread deployment of AI agents in high-stakes financial environments.
The integration of AI agents into core financial infrastructure marks a transition from AI as a tool to AI as an autonomous actor in commerce. By these agents directly into payment and treasury operations, Ant International is creating a standardized environment for agentic transactions, which is essential for scaling AI-driven commerce globally.
The introduction of specific security frameworks like KYA and AgentSafePay addresses a critical barrier to adoption: liability and trust. By providing indemnity schemes and verification standards, the company is attempting to mitigate the risks associated with autonomous decision-making in financial contexts, which could encourage broader enterprise adoption of AI agents.
The partnership with Visa and Mastercard to establish global agent verification standards through the Agentic Mobile Protocol has the potential to set industry-wide norms for how AI agents are authenticated and authorized in financial networks. This could influence how other fintech providers and banks approach the integration of AI agents into their systems.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').What is the most accurate way to describe what AI Agents can do today?
What to watch next
Monitor the rollout timeline through winter and the adoption rates of the new agent-specific business accounts. Watch for the implementation of the Agentic Mobile Protocol with Visa and Mastercard, which aims to set global standards for agent verification. Additionally, track the real-world performance of the FalconTST forecasting model and the effectiveness of the AgentSafePay indemnity scheme in mitigating losses from agent errors.
The completion of the global rollout through winter will determine the initial scale of adoption. Monitoring the number of merchants and consumers actively using the new AI agent features will provide insight into the practical utility of these tools.
The development and adoption of the Agentic Mobile Protocol with Visa and Mastercard is a key indicator of industry standardization. If this protocol becomes widely adopted, it could become the de facto standard for agent verification in global payments.
The performance of the FalconTST model in reducing hedging costs and the effectiveness of AgentSafePay in covering losses from agent errors will be crucial for validating the economic and security benefits of the upgrade. Independent verification of these claims will be important for assessing the true impact of the initiative.